Characterization of Canarian cheeses applying discriminant analysis to physico-chemical variables

Abstract

Discriminant analysis was applied to 84 artisan and commercial cheese samples from the seven Canary Islands (Spain), determined by different physico-chemical variables (SNT, NPN, NNH3, a(w), pH, moisture, NaCl, fat). All of them were selected to compute discriminant functions according to the method of Fatti and Hawkins. The generalized square distances among classes and the posterior probability for each cheese to belong to a class were calculated. According to discriminant analysis three of the seven classes considered had 100% well-classified cheeses, three had 91.67% and one had 83.33%.